[Principal component analysis on ultrasound indexes of schistosomiasis and the assessment on prevalence rate].

Zhonghua Liu Xing Bing Xue Za Zhi

The Department of Epidemiology, School of Public Health, Fudan University, Shanghai 200032, China.

Published: September 2003

Objective: To explore the synthetical index for diagnosing schistosomiasis with ultrasound and to assess the prevalence rate with the index.

Methods: Ultrasound indexes of schistosomiasis Japonicum were analyzed by principal component analysis, and the synthetical indexes were assessed by ROC curve.

Results: Among the abnormal rates of the 6 indexes, the lowest was 1.6% comparing with the highest of 59.5%. Significant difference was noficed among the abnormal rates (chi(2) = 631.1, P < 0.01). The individual correlation of the six indexes to each other as will as with age distribution was significant (P < 0.05). The three principal components reflected the degree of pathological changes on liver and spleen. The first principal component was the factor reflecting the degree of liver pathological changes, and the second and third principal components reflected the degree of pathological changes on spleen. The synthetical index D(1) = 0.047X(1) + 0.428X(2) + 1.247X(3) + 0.095X(4) + 0.002X(5) + 0.213X(6) - 12.837 was found by adding the three weight principal components, and it's area under the ROC curve was 0.957. When -1.70 was taken as the critical value, the abnormal rate of population was 66.3%, close to the resident's actual prevalence rate 66.9%.

Conclusion: Ultrasonography was considered as a method which could rapidly assessing the resident's prevalence rate in the endemic areas of schisitosomiasis Japonicum, and could also provide powerful information for development of strategy on chemotherapy.

Download full-text PDF

Source

Publication Analysis

Top Keywords

prevalence rate
12
principal components
12
pathological changes
12
component analysis
8
ultrasound indexes
8
indexes schistosomiasis
8
principal component
8
abnormal rates
8
components reflected
8
reflected degree
8

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!